The purpose of a routine bridge inspection is to assess the physical and functional condition of a bridge according to a regularly scheduled interval. The Federal Highway Administration (FHWA) requires these inspections to be conducted at least every 2 years. Inspectors use simple tools and visual inspection techniques to determine the conditions of both the elements of the bridge structure and the bridge overall. While in the field, the data is collected in the form of images and notes; after the field work is complete, inspectors need to generate a report based on these data to document their findings. The report generation process includes several tasks: (1) evaluating the condition rating of each bridge element according to FHWA Recording and Coding Guide for Structure Inventory and Appraisal of the Nation’s Bridges; and (2) updating and organizing the bridge inspection images for the report. Both of tasks are time-consuming. This study focuses on assisting with the latter task by developing an artificial intelligence (AI)-based method to rapidly organize bridge inspection images and generate a report. In this paper, an image organization schema based on the FHWA Recording and Coding Guide for the Structure Inventory and Appraisal of the Nation’s Bridges and the Manual for Bridge Element Inspection is described, and several convolutional neural network-based classifiers are trained with real inspection images collected in the field. Additionally, exchangeable image file (EXIF) information is automatically extracted to organize inspection images according to their time stamp. Finally, the Automated Bridge Image Reporting Tool (ABIRT) is described as a browser-based system built on the trained classifiers. Inspectors can directly upload images to this tool and rapidly obtain organized images and associated inspection report with the support of a computer which has an internet connection. The authors provide recommendations to inspectors for gathering future images to make the best use of this tool.

Researchers are embracing deep learning in various interdisciplinary research domains, recognizing undeniable benefits offered by deep neural networks. However, in order to meet the substantial computational demands for processing deep learning models, researchers extensively rely on cloud servers. Nevertheless, the shared nature of cloud servers encourages research labs and facilities to establish private clouds, ensuring exclusive access to computational resources and safeguarding data privacy. Creating a private cloud from bare metal presents challenges with existing provisioning solutions. These solutions not only come with a set of complex installation and configuration steps but are also limited to a constrained local Ethernet broadcast domain for network loading, which may pose unforeseen difficulties and risks for researchers who do not specialize in computing. To address these issues, this paper introduces EL2W, Extended Layer 2 services to Wide area networks (WAN), a novel approach we developed following Infrastructure-as-Code (IaC) principles. EL2W aims to help automate the system installation procedure by reducing the repetitive configurations and setups using Infrastructure-as-Code based scripts and codes. In addition, EL2W can securely expand an Ethernet network’s logical and functional extent beyond the current physical limitations of Ethernet layer 2 networks. We describe the implementation and architecture of a remote bare metal provisioning system built upon secure extended layer 2 networks. Experimental results demonstrate the capability of EL2W for establishing a secure layer 2 connection to provide essential bare metal provisioning services, as well as the effectiveness of a local proxy cache server to reduce the operating system loading time.

Modern information systems are built fron a complex composition of networks, infrastructure, devices, services, and applications, interconnected by data flows that are often private and financially sensitive. The 5G networks, which can create hyperlocalized services, have highlighted many of the deficiencies of current practices in use today to create and operate information systems. Emerging cloud computing techniques, such as Infrastructure-as-Code (IaC) and elastic computing, offer a path for a future re-imagining of how we create, deploy, secure, operate, and retire information systems. In this paper, we articulate the position that a comprehensive new approach is needed for all OSI layers from layer 2 up to applications that are built on underlying principles that include reproducibility, continuous integration/continuous delivery, auditability, and versioning. There are obvious needs to redesign and optimize the protocols from the network layer to the application layer. Our vision seeks to augment existing Cloud Computing and Networking solutions with support for multiple cloud infrastructures and seamless integration of cloud-based microservices. To address these issues, we propose an approach named Open Infrastructure as Code (OpenIaC), which is an attempt to provide a common open forum to integrate and build on advances in cloud computing and blockchain to address the needs of modern information architectures. The main mission of our OpenIaC approach is to provide services based on the principles of Zero Trust Architecture (ZTA) among the federation of connected resources based on Decentralized Identity (DID). Our objectives include the creation of an open-source hub with fine-grained access control for an open and connected infrastructure of shared resources (sensing, storage, computing, 3D printing, etc.) managed by blockchains and federations. Our proposed approach has the potential to provide a path for developing new platforms, business models, and a modernized information ecosystem necessary for 5G networks.

The workforce for research computing, cyberinfrastructure, and data analytics is a complex global ecosystem comprised of workers across academia, national laboratories, and industry. To explore the underlying factors that affect the growth and vitality of this workforce ecosystem, we conducted an NSF funded virtual workshop during the third quarter of 2020 attended by 100 participants. The workshop identified challenges affecting the workforce pipeline and ecosystem and generated recommendations to help address these challenges. This paper provides a summary of the workshop, challenges, and recommendations.

After hazard events, large numbers of images are collected by reconnaissance teams to document the post-event state of structures, and to assess their performance and improve design procedures and codes. The majority of these data are captured as images and manually labeled. This highly repetitive task requires considerable domain expertise and time. Advances in deep learning have enabled researchers to rapidly classify reconnaissance images. Thus far, these classification methods are limited to a simple classification schema in which the classes are all either mutually exclusive or independent. To date, an efficient classification system of a complex schema containing many classes arranged in a multi-level hierarchical structure is not available to support earthquake reconnaissance. To address this gap, this paper introduces a comprehensive classification schema and a multi-output deep convolutional neural network (DCNN) model for rapid postearthquake image classification. In contrast to past work, herein a single multi-output DCNN classification model with a hierarchy-aware prediction was trained to enable the rapid organization of images. The performance of the proposed multi-output model was validated through comparisons with multi-label and multi-class models using an F1-score. As result, the multi-output model outperformed other models. Then, the multi-output model was deployed to a web-based platform called the Automated Reconnaissance Image Organizer, which can be used to easily organize earthquake reconnaissance images. (C) 2022 American Society of Civil Engineers.

There has been tremendous growth in the amount of visual data available on the Internet in recent years. One type of visual data of particular interest is produced by network cameras providing real-time views. Millions of network cameras around the world continuously stream data to viewers connected to the Internet. This data may be used by a wide variety of applications such as enhancing public safety, urban planning, emergency response, and traffic management which are computationally intensive. Analyzing this data requires significant amounts of computational resources. Cloud computing can be a preferred solution for meeting the resource requirements for analyzing these data. There are many options when selecting cloud instances (amounts of memory, number of cores, locations, etc.). Inefficient provisioning of cloud resources may become costly in pay-per-use cloud computing. This paper presents a method to select cloud instances in order to meet the performance requirements for visual data analysis at a lower cost. We measure the frame rates when analyzing the data using different computer vision methods and model the relationships between frame rates and resource utilizations. We formulate the problem of managing cloud resources as a Variable Size Bin Packing Problem and use a heuristic solution. Experiments using Amazon EC2 validate the model and demonstrate that the proposed solution can reduce the cost up to 62 percent while meeting the performance requirements.

New usage patterns of computing for research have emerged that rely on the availability of flexible, elastic, and highly specialized services, that may not be well suited to traditional batch HPC. A new approach that updates and evolves the research computing ecosystem is needed to respond to these needs. This new model, a Kubernetes-based "Community Composable Platform", builds upon Purdue's Community Cluster program to provide cost effective, highly responsive, and customizable composable computing solutions for domain science and education in a variety of communities.

High performance computing systems comprised of hundreds or thousands of computational nodes can generate a high volume of system log entries at a high data velocity. Analyzing these logs soon after they are generated is a significant challenge, due to the complexity of log messages, the speed at which they are produced, and the lack of a method to quickly map or categorize messages to meaningful sets. The impact of this problem is that it is not possible to comprehensively glean timely information from logs about the overall system or the health of individual nodes. In this paper, we address this problem through the development of a novel approach for system log analysis based on a markov random field (MRF) that can quickly categorize system log messages into multiple categories based on representative training examples provided by a user. We present a theoretical model of our approach, followed by an extensive evaluation of the accuracy and performance of the implementation of our model. We found that our MRF based approach can quickly categorize system log messages with a high degree of accuracy.

Reconnaissance teams are charged with collecting perishable data after a natural disaster. In the field, these engineers typically record their observations through images. Each team takes many views of both exterior and interior buildings and frequently collects associated metadata that reflect information represented in images, such as global positioning system (GPS) devices, structural drawings, timestamp, and measurements. Large quantities of images with a wide variety of contents are collected. The window of opportunity is short, and engineers need to provide accurate and rich descriptions of such images before the details are forgotten. In this paper, an automated approach is developed to organize and document such scientific information in an efficient and rapid manner. Deep convolutional neural network algorithms were successfully implemented to extract robust features of key visual contents in the images. A schema is designed based on the realistic needs of field teams examining buildings. A significant number of images collected from past earthquakes were used to train robust classifiers to automatically classify the images. The classifiers and associated schema were used to automatically generate individual reports for buildings.

Dynamic response of structures is a complex process that is not well-understood. Seismic design codes allow structural systems of buildings to behave inelastically during strong ground shaking. While hysteretic energy dissipation due to inelastic behavior is rather well-understood, there exist other energy dissipation mechanisms which are not as well-understood. Energy dissipation due to mechanisms other than material non-linearity are often modeled in the form of a single, velocity proportional damping mechanism with an equivalent viscous damping ratio. In seismic design and dynamic analysis of structures, this equivalent viscous damping ratio is generally taken to be constant (e.g., 2 or 5% of the critical) regardless the response is elastic or inelastic. Instead of making such a strong assumption about the viscous damping ratio, which may have large influence on the peak response levels, it is advisable to use an effective viscous damping estimate based on studying the actual response of real structures responding to dynamic loads. Use of such obtained viscous damping will allow extraction of restoring forces from dynamic force estimates. At low strain-rate levels, such as those observed during seismic response, restoring forces match the resistance that develops during quasi-static loading, a method to estimate effective viscous damping and restoring force empirically from dynamic response of a structure is presented. The method considers inelastic response explicitly, i.e., no linearization assumptions are made for the load-deformation behavior of the structure. The presented method is tested on several computational simulation models with various hysteretic behaviors and a preset constant viscous damping ratio to verify that the algorithm (1) estimates a damping ratio close to the value used in the simulations, and accordingly, (2) captures the hysteretic behavior accurately. The method is illustrated using data obtained from reinforced concrete test specimens subjected to design-level base excitations on an earthquake simulator.

Machine learning is an important applied research area in particle physics, beginning with applications to high-level physics analysis in the 1990s and 2000s, followed by an explosion of applications in particle and event identification and reconstruction in the 2010s. In this document we discuss promising future research and development areas in machine learning in particle physics with a roadmap for their implementation, software and hardware resource requirements, collaborative initiatives with the data science community, academia and industry, and training the particle physics community in data science. The main objective of the document is to connect and motivate these areas of research and development with the physics drivers of the High-Luminosity Large Hadron Collider and future neutrino experiments and identify the resource needs for their implementation. Additionally we identify areas where collaboration with external communities will be of great benefit.

In this chapter of the High Energy Physics Software Foundation Community Whitepaper, we discuss the current state of infrastructure, best practices, and ongoing developments in the area of data and software preservation in high energy physics. A re-framing of the motivation for preservation to enable re-use is presented. A series of research and development goals in software and other cyberinfrastructure that will aid in the enabling of reuse of particle physics analyses and production software are presented and discussed.

Network cameras, a type of surveillance cameras, generate real-time, versatile, and high quality video content that can be used for applications such as public safety and surveillance. Analyzing high frame rate video streams imposes heavy computing needs and significant loads to the network. High frame rates may not be essential for meeting the accuracy requirements of the analyses. For example, high frame rates may not be required to track cars inside a garage compared with cars on a highway. In this paper, we study object tracking and propose a method to automatically determine the necessary frame rate for videos in network cameras for object tracking and adapt to runtime conditions. We demonstrate that the frame rates can be reduced up to 80% based on accuracy constraints.

Resource optimization has always been a big challenge in modern data centers. The process of performing workload consolidation on a minimal number of physical machines is becoming more complex when these data centers began supporting containers in addition to virtual machines (VMs). With the increasing usage of containers with VMs in data centers, it becomes critical to address this problem from the container's point of view - that is to optimally allocate containers in the fewest number of physical hosts. Depending on the type of application workload or tasks, infrastructure providers may provision separate containers to handle each task. These tasks may have different resource demands, such as: some of these tasks are CPU intensive, some memory intensive, some I/O intensive and some may be network intensive. Also, the physical machines in the data center are heterogeneous i.e. the hardware configuration (resource capacity) of these machines may differ from each other. Hence, the challenge is to consolidate all the active containers with different resource requirements on the minimum number of physical machines that are not even. We formulate a multi-resource bin packing problem and propose a Deep Learning technique called Fit-for-Packing to place a near-optimal number of containers on a physical machine. Experimental results show that our model achieves an average training accuracy of 82.01% and an average testing accuracy of 82.93%.

Indoor localization can provide a number of different services such as location-aware advertisement, indoor navigation and automating different appliances based on the user location. A number of different techniques such as time-difference-of-arrival, angle-of-arrival, time-of-flight, and received signal strength indicator (RSSI) have been used to provide Location Based Services (LBS). RSSI is one of the widely used methods as it is cost efficient and easy to implement. However, RSSI's performance is limited by multipath fading and indoor noise. Particle Filter (PF) is an accurate Bayesian Filtering algorithm that can improve the performance of RSSI-based indoor localization. However, PF is not able to satisfy the high accuracy requirement (possibly 10cm) of indoor localization. In this paper, we present Particle Filter-Extended Kalman Filter (PFEKF) cascaded algorithm that combines PF and EKF in series to reduce the impact of multipath effects and noise on the RSSI. Our experimental results show that PFEKF improves the localization accuracy by 31.3% and 33.9% in 3D and 2D environments respectively when compared with using only a PF.

Indoor localization and Location Based Services (LBS) can greatly benefit from the widescale proliferation of communication devices. The basic requirements of a system that can provide the aforementioned services are energy efficiency, scalability, lower costs, wide reception range, high localization accuracy and availability. Different technologies such as WiFi, UWB, RFID have been leveraged to provide LBS and Proximity Based Services (PBS), however they do not meet the aforementioned requirements. Apple's Bluetooth Low Energy (BLE) based iBeacon solution primarily intends to provide Proximity Based Services (PBS). However, it suffers from poor proximity detection accuracy due to its reliance on Received Signal Strength Indicator (RSSI) that is prone to multipath fading and drastic fluctuations in the indoor environment. Therefore, in this paper, we present our iBeacon based accurate proximity and indoor localization system. Our two algorithms Server-Side Running Average (SRA) and Server-Side Kalman Filter (SKF) improve the proximity detection accuracy of iBeacons by 29% and 32% respectively, when compared with Apple's current moving average based approach. We also present our novel cascaded Kalman Filter-Particle Filter (KFPF) algorithm for indoor localization. Our cascaded filter approach uses a Kalman Filter (KF) to reduce the RSSI fluctuation and then inputs the filtered RSSI values into a Particle Filter (PF) to improve the accuracy of indoor localization. Our experimental results, obtained through experiments in a space replicating real-world scenario, show that our cascaded filter approach outperforms the use of only PF by 28.16% and 25.59% in 2-Dimensional (2D) and 3-Dimensional (3D) environments respectively, and achieves a localization error as low as 0.70 meters in 2D environment and 0.947 meters in 3D environment.

Modern web/mobile applications have a highly variable user load from different geographical locations that depends upon the time-of-day, flash crowds, and growth rate (high or low) of user load. Irrespective of server hosting location of the applications, users from different geographical regions need highly responsive applications because bad performance results in unhappy customers. A recent survey suggests that mobile applications lose 34% of their users due to poor response time. To enhance the user experience of applications, we propose Global Live Migration (GLM) of VMs, which allows VMs to be migrated between different geographical regions over Wide Area Networks (WANs). Our experimental setup consists of Software-defined Network (SDN) enabled networking infrastructure because of its extensive control over network traffic routing. Our objective is to place the VMs on sites closest to the region with a major population of respective application users. However, there are high chances of congestion in inter-site links if a random migration sequence is followed while performing multiple VM migrations that may result in migration failures. We formulate a Mixed Integer Programming (MIP), and solve it using a Lagrangian relaxation based subgradient technique. We also propose a feedback-based control system with self-tuning regulators to improve the migration sequence regularly. Our extensive simulations show that our proposed solution achieves a minimum improvement of 38.09% compared to the Best Fit algorithm.

A select number of scientific communities have been quite successful in evolving the culture within their community to encourage publishing and to provide resources for re-using well-documented data. These data have great potential for analysis and knowledge generation beyond the purposes for which they were collected and intended. However, there are still barriers in this process. To explore this problem, we have developed a prototype tool: the Experiment Dashboard (ED), with the objective of demonstrating the ability and potential of enabling automated data ingestion from typical research laboratories. This innovative prototype was developed to create a novel system and artifact to explore the possibilities of allowing researchers in laboratories across the nation to link their data acquisition systems directly to structured data repositories for data and metadata ingestion. The prototype functions with commonly used data acquisition software at the data source and the HUBzero scientific gateway at the data sink. ED can be set up with minimal effort and expertise. In this paper, we describe the motivation and purposes for the prototype, the architecture we devised and functionality of this tool, and provide a demonstration of the tool for optical measurements in a structural engineering laboratory. The goal of this paper is to articulate and show through our prototype a vision for future cyberinfrastructure for empirical disciplines that rely on the rapid collection, analysis, and dissemination of valuable experimental data. We also discuss lessons learned that may be useful for others seeking to solve similar problems. (C) 2017 Elsevier Ltd. All rights reserved.

High performance computing (HPC) systems frequently suffer errors and failures from hardware components that negatively impact the performance of jobs run on these systems. We analyzed system logs from two HPC systems at Purdue University and created statistical models for memory and hard disk errors. We created a small-scale error injection testbed-using a customized QEMU build, libvirt, and Python-that could be used by HPC application programmers to test and debug their programs in a faulty environment so that programmers can write more robust and resilient programs before deploying them on an actual HPC system.

Proximity-based Services (PBS) require high detection accuracy, energy efficiency, wide reception range, low cost and availability. However, most existing technologies cannot satisfy all these requirements. Apple's Bluetooth Low Energy (BLE), named iBeacon, has emerged as a leading candidate in this domain and has become an almost industry standard for PBS. However, it has several limitations. It suffers from poor proximity detection accuracy due to its reliance on Received Signal Strength Indicator (RSSI). To improve proximity detection accuracy of iBeacons, we present two algorithms that address the inherent flaws in iBeacon's current proximity detection approach. Our first algorithm, Server-side Running Average (SRA), uses the path-loss model-based estimated distance for proximity classification. Our second algorithm, Server-side Kalman Filter (SKF), uses a Kalman filter in conjunction with SRA. Our experimental results show that SRA and SKF perform better than the current moving average approach utilized by iBeacons. SRA results in about a 29% improvement while SKF results in about a 32% improvement over the current approach in proximity detection accuracy.